What Are Intelligent Vehicles?

Intelligent vehicles are ground, air, or water vehicles equipped with sensing, computing, and communication systems that enable them to perceive their environment, reason about their situation, and perform driving or navigation tasks with varying degrees of autonomy. The spectrum ranges from production vehicles with driver assistance features such as automatic emergency braking to fully autonomous systems capable of operating without any human input across a defined set of conditions. Intelligent vehicle technology draws on disciplines including robotics, computer vision, artificial intelligence, communications, and control systems engineering.

Sensing: LIDAR, Cameras, and Sensor Fusion

Perception is the foundation of intelligent vehicle operation. A vehicle must accurately detect and classify objects in its surroundings, estimate their positions and velocities, and predict their future behavior to make safe navigation decisions.

LIDAR (light detection and ranging) is a primary sensor for autonomous vehicles, emitting pulsed laser light and measuring return times to construct precise three-dimensional point clouds of the environment. Mechanical spinning LIDAR systems scan 360 degrees and generate millions of points per second, while solid-state LIDAR variants use no rotating parts, offering improved reliability and lower cost at the expense of field of view. LIDAR provides accurate depth measurements in adverse lighting conditions but is affected by heavy precipitation and is relatively expensive compared to cameras.

Cameras are the richest source of semantic information, capturing texture, color, and scene context that LIDAR point clouds lack. Computer vision algorithms trained on large labeled datasets detect lane markings, traffic signs, pedestrians, and other vehicles. Camera-based systems depend on illumination and are challenged by glare, rain, and low-contrast conditions.

Radar sensors complement cameras and LIDAR with reliable velocity measurement through the Doppler effect and robust performance through fog, rain, and darkness. Millimeter-wave radar at 77 GHz is standard in production vehicles for adaptive cruise control and automatic emergency braking.

Sensor fusion combines data from multiple modalities to produce a unified environmental model that is more accurate and robust than any single sensor. Fusion algorithms, typically implemented using Kalman filters, particle filters, or learned neural approaches, weight each sensor's contribution based on its uncertainty and current operating conditions. The IEEE Intelligent Vehicles Symposium is the leading international conference for research on sensing, perception, and planning for intelligent vehicles.

Autonomous Driving and Path Planning

The SAE International J3016 standard defines six levels of driving automation from Level 0 (no automation) to Level 5 (full automation under all conditions). Most deployed production systems operate at Level 2 (partial automation), where the driver must remain engaged, or Level 3 (conditional automation), where the system can handle certain environments but requests driver intervention when conditions exceed its capability. SAE International's J3016 taxonomy is the globally referenced framework for describing automation levels and is used by regulators, automakers, and media to communicate vehicle capability accurately.

Path planning translates the vehicle's destination and perceived environment into a safe, comfortable trajectory. Planning systems typically operate in two stages: route planning over road networks and motion planning at the vehicle scale. Motion planning must navigate among other vehicles, pedestrians, and static obstacles while respecting traffic laws, ride comfort limits, and energy efficiency goals. Sampling-based algorithms such as RRT* and optimization-based approaches such as model predictive control are widely used. Deep reinforcement learning is an area of active research for handling complex urban interactions.

Unmanned vehicles extend intelligent vehicle technology to military ground robots, agricultural equipment, underwater vehicles, and aerial drones, each with domain-specific sensor suites. The U.S. Army Research Laboratory funds research on autonomous ground vehicle navigation in unstructured off-road environments.

Applications

Intelligent vehicle technology serves a broad and growing range of transportation and operational applications:

• Passenger vehicle safety systems at SAE Level 1 and 2, including automatic emergency braking, adaptive cruise control, and lane centering, are credited with measurable reductions in rear-end and lane-departure crashes.
• Autonomous taxi and robotaxi services operate in geofenced urban areas using full sensor suites and high-definition map data to provide passenger transport without a human driver.
• Autonomous trucking applies Level 4 automation to highway long-haul operations between distribution hubs where conditions are more predictable than in dense urban environments.
• Agricultural autonomous vehicles use GNSS-guided path following and precision application systems to plant, spray, and harvest crops with centimeter-level accuracy, reducing input costs and operator workload.
• Military unmanned ground vehicles perform logistics resupply, reconnaissance, and route clearance missions in environments too dangerous for human-crewed vehicles.
• Underground mining vehicles operate autonomously in GPS-denied tunnel environments using LIDAR-based simultaneous localization and mapping (SLAM) to haul ore and remove waste with higher utilization and improved worker safety.